dsplib rifft: can't get it working

Started by lolo_mk1 June 20, 2005

I've read past posts on this topic but I can't make the dsplib
inverse fft work properly.

I put it in a 360Hz sine wave @ 8kHz sampling rate. I get the
desired spike in the magnitude spectrum via

cbrev(*in, *out, FFTORDER/2);
rfft1024(*out, 1);
//my little magnituding finding function goes here

so I'm sure thats working. then...

cbrev(*out, *in, FFTORDER/2);
rifft1024(*in, 1);

the resulting amp/time graph is completely distorted with very high
frequency components introduced.

I checked and I do have the latest dsplib files. I've also tried
applying a hann window on the initial data to avoid spectral
widening. in and out are also aligned to 2048.

Can anyone see whats wrong here??